This electronic supplement includes other pictures and descriptions of surface effects of the Dôbi earthquake sequence, which we observed in the southeastern part of the Dôbi graben, in the Galafi-Kadda Hawli region.
Figure S1 shows northwest-looking view of large sand fountain in Area 3 (1-2 m across, location on Figs. B2 & B3 in appendix B).
Figure S2 is north-looking view of one spectacular open fissure in Area 5 (location on Fig. B8 in appendix B). Alluvial fan between F1 and F2 is visible in background. Notebook gives scale.
In February 1993, we also found open fissures across the sedimentary floor of the Makεanni half-graben, 2 km northwest of the Kadda Hawli half-graben (see Fig. 4b in the paper). The fissures were exposed at the base of, and roughly parallel to, a third order, N130°E-trending fault scarp. They were almost filled with sediment (Fig. S3). In April 1994, that part of the graben floor was flooded and we could not map this fissure set. However, about one kilometer west of it, we found other fissures that had gone unnoticed the year before, even though some crosscut the asphalted road. They lie just west of a third order fault, which cuts the block downthrown by the second order fault that formed the half-graben.
Based on their freshness, the fissures can be divided in two sets. The most weathered fissures, with degraded edges, of the first set were almost filled with sediments and grass, resembling those seen in 1993. They covered an area four times wider than that of the second set. The fresh fissures of the second set, which we mapped accurately with the total station (~ 300 data points) (Fig. S4a & S5), covered an area only 150x80 m2 wide. They were several tens of meters long, ~ 30 cm wide and a few meters deep at most. Minimal weathering of their edges suggests that they suffered only a few rainy seasons, if any. Even the smallest fissures (a few mm to a few cm wide) exhibited sharp angular edges. As a whole, these fissures again show a peculiar pattern, with two comparable subsets, crosscutting or connecting one another at right angles. This pattern resembles that of the orthogonal fissures in Area 7 of Kadda Hawli. The central fissure (Fm) connects to sub-orthogonal fissures (fm1 to fm4, Fig. S4a & S5) with trends averaging N90°-120°E, and N0-40°E, respectively. Such a geometry, which shows no strike-slip motion, implies roughly equal values of σ3 and σ2, oriented respectively N15°E and N110°E.
Because the freshest fissures of the Makεanni depression, which we did not notice in February 1993, might not have withstood one rainy season, it is possible that they were caused by the M~5.5 16/03/1993 earthquake, located by the Geophysical Observatory of Arta (CERD-IPGP) 20 km west of Makεanni. This shock probably occurred on D5. Small stress changes induced by the 16/03/1993 event might thus have caused blind ruptures created by the 1989 Dôbi earthquake sequence to propagate to the surface. In the same small depression, the older group of ruptures may have been the result of the Dôbi earthquakes. No other significant seismic sequence, except the much older and distant Serdo (1969) sequence, is known to have occurred around the Dôbi region. In the Makεanni half-graben, close to the wide and flood-prone plain of Kadda Dôbi, it is easy to envisage that older fissures might have been filled and eroded by spring floods between their formation (1989) and discovery (1993-1994).
In April 1994, we mapped open fissures in the εounda Hawli half-graben, between Kadda Hawli and Makεanni (see Fig. 4b in the paper). The fissure array covered an area of 140x100 m2 (~ 300 data points), with a mean trend of N150°E (Fig. S4b). These fissures lie on a slightly elevated part of the graben floor, which is thus mostly preserved from flooding. They cut calcareous lacustrine sediments covered by a thin veneer of basaltic gravel (colluvium derived from the tilted surface of block B). The fissures edges are more weathered than those of the freshest Makεanni ruptures. Their widths reach 30 cm (Fig. S6). The "trunk" of the fissure network consists of a system longer than 120 m (Fo, F'o and F"o) the trend of which veers counterclockwise 220° from N10°E to N150°E, then back to N120°E and N90°, from north to south (Fig. S4b). Several shorter fissures trending ~ N90°E in the north (fo1, fo2, fo3) and N10-20°E in the south (fo4, fo5), are connected with the main fissure system, sometimes at right angles. The widest fissures (e.g., F"o) together with those having a small vertical throw (~ 5 cm e.g., F'o) strike on average between N105° and N130°E. At a more detailed level, the main N150°E-trending section of Fo is an array of smaller right-stepping fissures, suggesting a left-lateral component of slip. Similarly, fo4 and fo5, which are connected to F'o are composed of right-lateral échelons of small fissures, a few mm to 1 cm wide, striking N0-10°E. This could reflect slight counter-clockwise rotation of the block located in between.
Fig. S3. West-looking view of filled fissure in eastern part of Makεanni half-graben (February 1993).
Fig. S4. Map of surface ruptures in (a) Makεanni and (b) εounda Hawli areas measured with total station (in April 1994).
Fig. S5. West-looking view of most recent open cracks measured with total station in central part of Makεanni depression (April 1994, location on Fig. S4a). Footprint in foreground gives scale. D1 cumulative scarp is visible in background.
Fig. S6. West-looking view of principal surface rupture F'o in εounda Hawli half-graben (1994, location on Fig. S4b). Note tilted blocks and kink of D1 in background and small open cracks in foreground.
[ Back ]
